TENCENT(HK:00700) Mei Ri Jing Ji Xin Wen·2025-10-28 14:39Core Insights - The integration of large models with enterprise knowledge bases is becoming essential for the financial industry to address challenges such as knowledge silos and high compliance costs [2][3][5] Group 1: Role of Knowledge Base - The enterprise knowledge base acts as a foundational infrastructure in the era of large models, where the large model serves as the "brain" and the knowledge base as the "textbook" [3] - The development of a "general model + specialized knowledge base" is akin to a "brain + textbook" model, requiring significant resources for training a versatile base model and equipping it with a comprehensive, real-time updated financial knowledge base [3][4] Group 2: Data Security Concerns - Financial institutions have heightened requirements for knowledge base security and compliance due to regulatory demands [5][6] - The construction of AI knowledge bases for financial enterprises necessitates specialized capabilities, including multi-level permission management, anti-leakage measures, and rigorous auditing capabilities [6] Group 3: Accuracy and Rigor in Knowledge Management - The financial sector demands high accuracy and rigor in knowledge management, as employees frequently interact with complex data and knowledge [7] - There is a growing emphasis on the accumulation of individual employee knowledge within financial firms, leading to the concept of "personal knowledge bases" [7] Group 4: Addressing "Hallucination" Issues - The phenomenon of "hallucination," where AI models generate incorrect responses, is prevalent, and addressing this requires a focus on several key areas [9] - The company has implemented various strategies to enhance the accuracy of its AI knowledge base, achieving an accuracy rate exceeding 92% through advanced document understanding and retrieval techniques [10][11][12]